Predicting success on conditional release for insanity acquittees: regionalized versus
nonregionalized hospital patients.
C. Tellefsen, M. I. Cohen, S. B. Silver and C. Dougherty,
Bull. Amer. Acad. Psychiatry & the Law
20(1): 87-100, 1992.
This research compared the outcomes of two cohorts of insanity acquittees: one group was treated
solely in the maximum security state forensic hospital before their release to the community
(nonregionalized) and the other group was treated at the state forensic hospital and transferred for
further treatment at less secure state regional hospitals (regionalized). This research describes the
outcome of a group of insanity acquittees (regionalized patients) never previously studied. The
applicability of a prediction model based on earlier research of insanity acquittees was tested on the
patients. Findings on four outcome indicators are reported: rearrests within five years after release,
overall functioning in the community five years after release, rehospitalizations for mental illness,
and successful completion of the terms of the five-year conditional release (nonrevocation).
Discriminant analysis was performed on the four outcome variables. The model was found to
accurately predict the four types of outcome from 69 percent to 94 percent accurately for the
nonregionalized insanity acquittees and from 87.5 percent to 95.8 percent for the regionalized
patients. This model is currently being adapted to classify patients into potential high- and low-risk
groups at the time of conditional release for the purpose of determining the intensity of outpatient